Improvement of Texture Image Retrieval Algorithm Based on Sparse Coding

Abstract

The demerits of the texture image retrieval algorithm based on sparse coding are found expression in the low recall and inconspicuous serial priority of qualified images. Several methods are presented in this paper to improve the performance of the image retrieval algorithm. Firstly, Brodatz texture image filter basis function is used for processing the texture images, thereafter, the kurtosis is added to generate the eigenvector, finally, the joint-scale filter basis is utilized to advance the filter effect. Experimental results indicate that the performance of the proposed methods is positive.